NCJ Number
113904
Journal
Journal of Quantitative Criminology Volume: 4 Issue: 2 Dated: (June 1988) Pages: 173-186
Date Published
1988
Length
14 pages
Annotation
This paper reviews and develops the range of stochastic models that take account of variations in the rate of committing offenses (lambda) that take place during an individual offender's career, variations in lambda between different types of offenders, and both forms of variation simultaneously.
Abstract
These models are all within the general class of nonhomogenous Poisson processes. They assume that within a period of followup or within an individual's criminal career, the number of occurrences of an event such as an arrest as well as the exact times at which these events occur, are determined by a chance process. The models based on this assumption have an important mathematical simplicity in that they belong to the class of Markov processes. Together these models provide a flexible and powerful tool for studying criminal careers. They also make a major contribution toward addressing a specific recommendation of the National Research Council Panel on Research on Criminal Careers. Figures and 19 references. (Author abstract modified)